Supplemental material for "On the Use of Self-Supervised Speech Representations in Spontaneous Speech Synthesis"
Anonymous submission to SSW 2023

Each of the two folders 'audio_tsgd' and 'audio_tcc' contains TTS samples from 18 systems trained (details in paper).
First 3 randomly picked utterances are included for all 18 systems in each corpus.

The format of each wav file is: m[SSL_model]_l[SSL_layer]_s[sample_id].wav

SSL_model = 
0: d2v_base_ssl
1: d2v_base_asr
2: w2v2_base_ssl
3: w2v2_base_asr
4: whisper_small
5: wavlm_bp

SSL_layer is 0-index version of 1-index layers of our experiment: 
5 (layer 6), 8 (layer 9), 11 (layer 12)